Shopify Redesign vs Iterate: A Decision Framework | PMD

Every founder running a £5M to £200M Shopify brand will, at some point this year, sit in a Monday morning meeting and hear the phrase: we need a redesign. It usually arrives after a bad week, a board deck, a competitor relaunch, or a new head of marketing settling in. It feels decisive. It feels like progress. It is, more often than not, the wrong call.

And just as often we see the opposite mistake. Brands grind on micro-tests for eighteen months on a theme that fundamentally cannot support the experience they are trying to sell, then wonder why CVR refuses to budge past 1.4%. They are iterating on something that needs to be rebuilt.

This is the question we get asked more than any other at PM Digital Design: should we redesign, or should we iterate? The honest answer is that the question itself is wrong. The better question is what is actually broken, and the answer to that determines the action.

Below is the 5-factor decision framework we use internally before recommending one or the other to a client. It is opinionated. It is designed to stop you spending six months and £200K on a redesign you did not need, and it is equally designed to stop you A/B testing your way around a problem that requires a rebuild.

The Shopify redesign vs iterate question is the wrong question

If you ask redesign or iterate? the answer is binary and emotional. If you ask where is the conversion leak and what is the minimum intervention that fixes it? the answer is structural and testable.

Most agencies will sell you whichever they are best at delivering. Design shops sell redesigns. Optimisation shops sell tests. PMD sits in the middle because we sell profit optimisation, not deliverables, and that means the same project might look like a six-week iteration sprint on one brand and a full theme rebuild on another. The brief should follow the diagnosis, not the other way round.

Before we touch a single Figma file or open a single Shopify theme file, we run five diagnostic factors. Each one returns a score. The scores point you to one of three outcomes: iterate, partial rebuild, or full redesign. We will go through each factor with the actual questions we ask, the data we look at, and the threshold that tips us one way or the other.

Factor 1: Is the conversion problem distributed or concentrated?

Open your funnel report. Look at landing CVR by page, add-to-cart rate by template, checkout completion by step. Where is the bleed?

If you find one page (typically a hero PDP or a single landing page) doing 60% of the volume and underperforming the rest of the site, your problem is concentrated. A redesign is overkill. You need a focused page-level intervention. This is iteration territory, and a small dedicated CRO programme will usually beat a redesign here because you are surgically fixing the one thing that matters.

If you find that conversion is broadly underperforming benchmarks across category pages, PDPs, cart, and checkout, your problem is distributed. Distributed problems rarely respond to surgical fixes. They are usually symptoms of structural failures, navigation logic, theme architecture, brand inconsistency, or a fundamentally mismatched information hierarchy. That is redesign territory.

The diagnostic: pull a 90-day cohort, segment by template type, compare against the CRO audit benchmarks we publish. If three or more templates are underperforming by more than 25%, the problem is distributed.

Factor 2: Has the brand positioning materially shifted in the last 18 months?

If your founder, your CMO, or your board have meaningfully changed who you are selling to, what promise you are making, or which category you are competing in within the last 18 months, your site is almost certainly lying to your customer. No amount of A/B testing the buy button colour will fix a site that is positioned for the wrong audience.

The questions to ask: has our pricing tier moved (premium to mass, or mass to premium)? Has the hero product category changed? Has the brand voice been formally rewritten? Have we acquired or launched a sub-brand that now sits awkwardly inside the same theme?

If you answered yes to any of those, the cost of not redesigning is steeper than it looks. You are paying for paid traffic against a positioning your site no longer delivers on. The bounce rate is hidden tax.

If you answered no, you are dealing with an execution problem, not a positioning problem, and iteration will almost always win.

"A redesign solves a positioning problem. An iteration sprint solves an execution problem. Confusing the two is the most expensive mistake we see at this stage of growth."

Factor 3: Is the theme codebase a blocker for test velocity?

This is the factor most founders never check, and it is the one that secretly kills more CRO programmes than any other.

Ask your developer (or your agency) three questions. One: how long does it take to push a new section variant live behind a feature flag? Two: can we run two concurrent A/B tests on the PDP without one breaking the other? Three: is the theme on Online Store 2.0 with proper section blocks, or is it a legacy 1.x theme with hardcoded sections?

If the honest answers are three weeks, no, and 1.x with everything hardcoded, your codebase is the blocker. You can iterate forever and barely move the number, because every test is expensive to ship and even more expensive to learn from. We have seen brands ship two tests a quarter on this kind of theme. Two. A quarter.

A modern Shopify theme architecture should let you ship two to four tests a week. If yours cannot, the rebuild pays for itself inside one year purely on test velocity, before you count any direct CVR lift. This is the case where a redesign is, paradoxically, the cheapest path to more iteration. Our custom Shopify build practice exists precisely for this scenario.

Factor 4: What is the cost of being wrong?

A redesign that drops CVR by 15% on a £3M brand is annoying. The same drop on a £30M brand is a P&L event that ends careers. The cost of being wrong scales with revenue, and the decision framework has to scale with it.

Three sub-questions matter here. How much of your revenue comes from organic and direct (which depends on the existing site holding its SEO and brand equity)? How fragile is your paid attribution (which gets disrupted whenever you change URLs or page structures)? How resilient is your customer base to a brand visual shift?

If you are heavily paid-acquisition led, low organic dependency, and your customer base is broadly transactional, the cost of being wrong on a redesign is manageable. You can afford to make a bold call. If you are organic-heavy, brand-loyal, and your existing customers actively refer (think the typical £20M+ subscription brand), the cost of being wrong is significantly higher and you should default to partial rebuild rather than full redesign. Refactor the theme, rebuild the funnel pages, leave the brand visual identity broadly intact, and ship in waves.

We dig into the financial framing of this in more detail in our LTV to CAC diagnostic toolkit, which is the same framework we walk CFOs through before signing off a rebuild budget.

Factor 5: Can you ship the same outcome with a 30-day landing page rebuild?

This is the question that, more than any other, has saved our clients millions of pounds in unnecessary redesign budget.

Before you commit to a full redesign, ask: if I rebuilt just the top three landing pages and the PDP for my hero product, would that solve 80% of the problem? Be brutally honest about the answer.

For most brands at the £5M to £20M stage running heavy paid acquisition, the answer is yes. The home page barely matters because cold traffic does not see it. The category pages barely matter because intent is bottom-funnel. What matters is the post-click experience on the pages that take ad spend, and that can usually be rebuilt in 30 days as a parallel project without touching the rest of the theme.

This is, frankly, our most-recommended path at this growth stage. It is also the path most agencies will quietly steer you away from because it is a smaller invoice. We walk through it in detail in our cold traffic landing page framework.

The case study: when a full redesign was the right call

Rory's Travel Club came to us on a legacy theme that could not support the membership-led experience they were trying to sell. The conversion problem was distributed across every template. The brand positioning had materially shifted from a transactional travel commerce site to a membership and community proposition. The codebase was the bottleneck for absolutely every test we proposed.

Three of the five factors landed firmly on the redesign side. The cost of being wrong was real, but the cost of not redesigning was higher. We migrated the store and shipped a custom theme. The full breakdown of how that 650% growth came together is in the Rory's Travel Club case study, and it remains one of our clearest illustrations of the framework above being applied correctly.

The counter-example, equally instructive, is the £14M apparel brand we worked with last year. Same conversation in the boardroom (we need a redesign), but when we ran the five factors, the score came back firmly on iteration. Their CVR problem was concentrated on three landing pages. Positioning was stable. The theme was Online Store 2.0 and perfectly capable. We delivered a 30-day landing page rebuild and a 12-week CRO sprint and beat their redesign business case on lift, on speed, and on cost.

The counter-view: when iteration is the more dangerous choice

The standard agency answer to the redesign question is just iterate, redesigns are risky. It sounds humble. It is often wrong.

Iteration is dangerous when your theme cannot support the tests you want to run. Iteration is dangerous when the positioning is no longer true and every winning test is just polishing a misaligned proposition. Iteration is dangerous when the compound drag of a tired site is killing margin every month you delay, and the cost of delay exceeds the cost of the rebuild.

The right question is not which is safer? The right question is which one matches what is actually broken? If the diagnosis points to a rebuild, iterating is the expensive choice. If it points to iteration, redesigning is the expensive choice. There is no universally safe answer.

Your Monday morning checklist

If you are sitting on a redesign decision this week, run through this before any agency conversation:

1. Pull the funnel data. Identify whether the CVR problem is concentrated or distributed. Three or more underperforming templates = distributed.
2. Audit your positioning. Has it materially shifted in 18 months? If yes, weight the score toward redesign.
3. Interview your developer. Ask the three test-velocity questions. If shipping a variant takes more than five business days, your theme is the blocker.
4. Quantify the cost of being wrong. What is the worst-case revenue drop you can absorb? Match the intervention to your tolerance.
5. Run the 30-day LP rebuild test. Ask if a parallel landing page rebuild would solve 80% of the problem. If yes, do that first.
6. Talk to a partner who does both. Most agencies sell one or the other. Find one that has shipped both and has no financial interest in the answer.

This is the same framework we use to scope every engagement at our agency, and it sits underneath the broader profit optimisation framework we apply across every client. If you want help running it on your own brand, book a 30-minute call with Paddy McLarnon and we will walk through it on the call. There is also a deeper learning path in the PMD CRO learning hub if you want to do the homework first.

FAQs

How long should a Shopify redesign take?

For a brand at £5M to £50M, a full redesign on a custom Shopify theme typically runs 12 to 16 weeks from kick-off to launch, assuming the brand work is done up front and content is ready. Anything faster usually means corners cut on accessibility, performance, or proper section architecture. Anything slower usually means scope creep, and you should ask why.

What is the typical cost of a Shopify redesign for a brand at our stage?

Pricing varies wildly depending on scope and agency, but a credible custom theme rebuild for a £5M to £30M Shopify brand will sit in the £60K to £200K range. If you are quoted less than that for a true custom build, ask hard questions about templating, performance budgets, and post-launch CRO support. If you are quoted significantly more, ask what you are getting for the additional spend beyond logo files and Figma pages.

How do I know if my theme is actually blocking test velocity?

Three signs: shipping a section variant takes more than a week, you cannot run two PDP tests concurrently without conflicts, and your developer talks about "hardcoded sections" rather than dynamic blocks. If two of those three are true, your theme is the bottleneck. Iteration will plateau.

Should we redesign before, during, or after a Shopify Plus migration?

If you are already migrating to Shopify Plus, fold the redesign into the migration. Doing them separately doubles the disruption and you lose the architectural benefit of redesigning on a modern foundation. The exception is if your team capacity cannot absorb both at once, in which case migrate first on a tidied-up version of the existing theme, then redesign in a second phase.

What is the safest first move if we cannot decide?

Rebuild your top one or two landing pages as a parallel project. It costs a fraction of a full redesign, it gives you a clean directional signal on whether the problem was design or execution, and it does not commit you to the bigger bet. We have seen brands answer the redesign question definitively from a 30-day landing page sprint alone.

Who should own the redesign vs iterate call internally?

Not the head of design, not the head of paid. The CMO or the founder, with diagnostic input from CRO, paid, and engineering. The reason: this is a P&L decision, not a craft decision. The people closest to the work tend to over-index on their own discipline. The decision needs someone with enough distance to weight all five factors honestly.